From Newsgroup: comp.lang.tcl

I have a function that returns today's date/time at 00:00:00, e.g., 2025-07-24T00:00:00, returned as seconds:

proc start_of_day {} {
clock scan "[clock format now -format %Y]-[clock format now \
-format %m]-[clock format now -format %d]" -format "%Y-%m-%d"
}

Is there a better way?
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl


Mark Summerfield <m.n.summerfield@gmail.com> posted:

I have a function that returns today's date/time at 00:00:00, e.g., 2025-07-24T00:00:00, returned as seconds:

proc start_of_day {} {
clock scan "[clock format now -format %Y]-[clock format now \
-format %m]-[clock format now -format %d]" -format "%Y-%m-%d"
}

Is there a better way?

clock scan [clock format now -format %Y-%m-%d]
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On 7/24/25 01:02, Mark Summerfield wrote:

Is there a better way?

What about [clock scan "today 00:00:00"] ?

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Mark Summerfield <m.n.summerfield@gmail.com> wrote:

I have a function that returns today's date/time at 00:00:00, e.g., 2025-07-24T00:00:00, returned as seconds:

proc start_of_day {} {
clock scan "[clock format now -format %Y]-[clock format now \
-format %m]-[clock format now -format %d]" -format "%Y-%m-%d"
}

Is there a better way?

proc start_of_day {} {
return [clock format [clock seconds] -format "%Y-%m-%dT00:00:00"]
}
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

Rich <rich@example.invalid> wrote:

Mark Summerfield <m.n.summerfield@gmail.com> wrote:

I have a function that returns today's date/time at 00:00:00, e.g.,
2025-07-24T00:00:00, returned as seconds:

proc start_of_day {} {
clock scan "[clock format now -format %Y]-[clock format now \
-format %m]-[clock format now -format %d]" -format "%Y-%m-%d"
}

Is there a better way?

proc start_of_day {} {
return [clock format [clock seconds] -format "%Y-%m-%dT00:00:00"]
}

Ah, missed the "returned as seconds".

proc start_of_day {} {
return [clock scan [clock format [clock seconds] -format "%Y-%m-%dT00:00:00"] -format %Y-%m-%dT%H:%M:%S]
}
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On 7/24/2025 4:02 AM, Mark Summerfield wrote:

I have a function that returns today's date/time at 00:00:00, e.g., 2025-07-24T00:00:00, returned as seconds:

proc start_of_day {} {
clock scan "[clock format now -format %Y]-[clock format now \
-format %m]-[clock format now -format %d]" -format "%Y-%m-%d"
}

Is there a better way?

I am curious why a better way? It doesn't seem that heavy or slow.

However, you can make it more "robust". It is possible that on a Dec 31,
2024, you will get 2024-Jan-01T00:00:00 because you are treating 3
different time points as "now". After the firs call for the year, the
clock may switch to the next year.


--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On 7/24/2025 12:48 PM, saito wrote:

However, you can make it more "robust". It is possible that on a Dec 31, 2024, you will get 2024-Jan-01T00:00:00 because you are treating 3
different time points as "now".  After the firs call for the year, the clock may switch to the next year.

Well, it is highly unlikely, with chances being slimmer thank winning
the mega lottery.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

saito <saitology9@gmail.com> wrote:

On 7/24/2025 12:48 PM, saito wrote:

However, you can make it more "robust". It is possible that on a Dec 31,
2024, you will get 2024-Jan-01T00:00:00 because you are treating 3
different time points as "now".  After the firs call for the year, the
clock may switch to the next year.

Well, it is highly unlikely, with chances being slimmer thank winning
the mega lottery.

Sadly, these "highly unlikely" items are just the things to come out of (seemingly) nowhere after the code's been running for years, and then,
one time, on Dec 31 the planets align "just right" and the process goes "boom".

Only by then, no one knows why, and possibly, the original author has
also moved on to elsewhere.

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On Thu, 24 Jul 2025 02:19:17 -0700, EvoTcl wrote:

On 7/24/25 01:02, Mark Summerfield wrote:

Is there a better way?

What about [clock scan "today 00:00:00"] ?

Thanks for all the replies; this is the one I used.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On 24/07/2025 11:19, EvoTcl wrote:

On 7/24/25 01:02, Mark Summerfield wrote:

Is there a better way?

What about [clock scan "today 00:00:00"] ?

This uses free-format scan, which the manual page says is deprecated.
The non-deprecated version would be: [clock scan 00:00:00 -format %T].
You can add the -base option if you want something other than today. But
-base defaults to now, so it is unnecessary for "today".


Schelte.

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: comp.lang.tcl

On Sat, 26 Jul 2025 22:28:31 +0200, Schelte wrote:

On 24/07/2025 11:19, EvoTcl wrote:

On 7/24/25 01:02, Mark Summerfield wrote:

Is there a better way?

What about [clock scan "today 00:00:00"] ?

This uses free-format scan, which the manual page says is deprecated.
The non-deprecated version would be: [clock scan 00:00:00 -format %T].
You can add the -base option if you want something other than today. But -base defaults to now, so it is unnecessary for "today".

Schelte.

Thanks, I'm now using your non-deprecated approach.
--- Synchronet 3.21a-Linux NewsLink 1.2